Skip to content
Agentmart
Submit a tool

setup

Install the AI-native SDLC loop into a product repository.

Skill by bibutikoley

IndexedMIT★ 3
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/bibutikoley-setup

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Install the AI-native SDLC loop into a product repository. Asks about CI platform (GitHub/GitLab/both), tracker, chat, monitoring, build/test/lint commands and protected paths, then scaffolds sdlc/, sdlc.config.json, REVIEW.md, CI workflows, evals, monitoring, MCP connectors and the CLAUDE.md section, idempotently. Use on /sdlc:setup or when asked to set up / adopt / install the SDLC workflow in a repo.

Use cases: testing, devops, productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/setup

What's inside

Skill name
setup
Description
Install the AI-native SDLC loop into a product repository. Asks about CI platform (GitHub/GitLab/both), tracker, chat, monitoring, build/test/lint commands and protected paths, then scaffolds sdlc/, sdlc.config.json, REVIEW.md, CI workflows, evals, monitoring, MCP connectors and the CLAUDE.md section, idempotently. Use on /sdlc:setup or when asked to set up / adopt / install the SDLC workflow in a repo.
License
MIT
Instructions
56 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# /sdlc:setup — one command for the platform engineer

Read `sdlc-loop`. This is a one-time task per repo. It must be safe to re-run: nothing existing is overwritten
without `--force`; conflicts are written beside the original as `<file>.sdlc-new`.

## 1. Discover (read-only)

- Confirm we are at a Git repo root. Detect: remote host (github.com / gitlab), default branch, existing
  `CLAUDE.md`, `Makefile`/`package.json`/`pyproject.toml`/`go.mod`/`pom.xml` to propose build/test/lint
  commands, existing `.github/workflows` or `.gitlab-ci.yml`, generated directories, migrations, infra.
- Existing `sdlc.config.json` means an upgrade: keep its values, only add missing keys.

## 2. Ask (only what discovery could not answer; batch the questions)

Platform · tracker (+ base URL, project key) · chat · monitoring · source of truth (`repo` recommended to
start, `linked` if a tracker already holds the record) · build/test/lint/format/run commands · protected paths
(propose from discovery) · which MCP connectors to wire (default: the platform's + tracker's + chat's +
monitoring's).

Ask about roles last, and separately: "Who holds each role, by git commit email? (product owner, engineer,
QA, tech lead, platform engineer, security/compliance, release manager, service owner, leadership — a person
can hold more than one, and any role you leave empty stays unenforced.)" Someone answering "not yet" is a
valid answer — it is fine to scaffold with `roles` empty and fill it in later.

## 3. Scaffold

Run `${CLAUDE_PLUGIN_ROOT}/scripts/scaffold.sh` with the answers as flags (see `--help`). It:
- writes `sdlc.config.json` (merged), `sdlc/` (README, lessons, templates, worked example), `REVIEW.md`,
  `evals/`, `monitoring/`, `.claude/settings.json` (permissions allow-list for the configured commands),
  `CODEOWNERS.example`, `managed-settings.example.json`;
- copies the CI files for the chosen platform(s) (`.github/workflows/sdlc-*.yml`, `.gitlab/sdlc/.gitlab-ci.sdlc.yml`);
- merges the chosen connectors into `.mcp.json`;
- appends the SDLC section to `CLAUDE.md` (creates a minimal one if absent) with the commands filled in;
- copies the policy skill templates into `.claude/skills/` **only if** the org has no policy skills yet.
Use `--dry-run` first and show the plan of files to the user.

`scaffold.sh` does not take role emails as flags. After it runs, edit `sdlc.config.json`'s `roles` object
directly with the answers from step 2 (each role is an array of emails; `roleCommands` ships with sane
defaults and rarely needs changing). **The moment any `roles.*` array is non-empty, the `role-gate` hook
starts enforcing it** for every command not in `roleCommands["*"]` — say this to the user before you write
the emails in, so it is not a surprise the next time someone runs an out-of-role command.

## 4. Finish with the human checklist (print it; do not do these yourself)

1. Set CI secrets: `ANTHROPIC_API_KEY` (or Bedrock/Vertex/Foundry variables), tracker/chat tokens named in `.mcp.json`.
2. Branch protection on the default branch: require PR, require code owner review, no direct pushes (agents included).
3. Create the `production` environment with required reviewers = release managers (GitHub) / protected environment (GitLab).
4. Fill in `.claude/skills/*/SKILL.md` policy templates with real policy text; each has an owner and a source link.
5. Add `CODEOWNERS` from the example. Add `sdlc.config.json` and `.claude/**` to code-owner review.
6. Regulated org: deploy `managed-settings.example.json` via MDM/admin console after tailoring the deny list.
7. Run `/sdlc:status`, then walk one small change through `/sdlc:intent` → `/sdlc:spec` → `/sdlc:plan` → `/sdlc:build`
   **before** filling in `roles` — once populated, the `role-gate` hook restricts each command to its role.
8. Fill in `sdlc.config.json`'s `roles` with the emails from step 2 when the team is ready to enforce it.
9. Read `docs/adoption-roadmap.md` for the recommended order of the remaining plays.

From bibutikoley/ai-sdlc-claude, MIT.

More from bibutikoley/ai-sdlc-claude

Signals

Stars
3
Forks
0
Open issues
0
Last commit (this folder)
Sep 5, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from bibutikoley

Cite this page

setup skill. Agent Mart. https://agentmart.dev/skills/bibutikoley-setup (last updated 2026-10-24)